Yes I have a brief one, sorry that it's not specifically related to Readlang!

You kind of stopped blogging as much when you joined Duolingo, what would you say were the biggest factors why? I'm assuming Duolingo weren't super restrictive on what you could post so that might not be the big reason, but I could be wrong!

Good question! There was a lot to do and I was kept busy just with internal communication, so didn't feel the urge to share externally so much. Also, getting outside attention didn't feel as important given my role at the company compared to when I was alone running my own project.

I actually remember asking them how restrictive they would be when it comes to sharing when I joined and they said they weren't as secretive as Apple, but that they weren't completely open about everything either. If I had had the urge to post anything I would have run it past people before publishing details about what our team was working on, and it would have probably been fine. But in the end, like I said, I was always busy and blogging fell by the wayside. (I was never a naturally prolific blogger anyway)

Thank you for answering!

The internal/external rift is so interesting and perfectly reasonable. Hard to spend time self editing after an exhausting day, and guess that's where places like Twitter really shined. Yet I wish that gap could be bridged somehow, it feels like many great posts with overlooked details are left unwritten because of the lack of external validation combined with internal hurdles.

Meandering now, but your blogs are enjoyable reads and that's the most envious quality I have of writers!

Without asking for numbers you don't want to share, how did the price you acquired it back for compared to the price they originally purchased it for?

Hmmm... I'm actually not sure if I can talk about the numbers from the original sale in 2017. But I'll say that the original sale was paid partly in stock options which vested over a few years and ended up being worth a lot. And the price of the buy back was much much less than that.

Funnily enough, in the original sale I negotiated for more cash and fewer options. Luckily for me they only moved a little bit in my direction since the options ended up being worth much more.

Not related to the buyback but more generally to Readlang, how did you select the pricing?

I've noticed a sort of bimodal distribution of prices in language companies with the sleek/modern looking ones centering around 5$/month, and more traditional companies going for 20$/month for seemingly the same service (so the latter not presented as a luxury alternative to the former). Wondering if you did any A/B testing or generally what your methodology was if you don't mind sharing.

Not a ton of methodology behind it!

I started at $10 per year, then gradually raised prices as the product got better, always keeping existing subscribers happy by grandfathering them. At one point it was $4.99 every 3 months, which is the plan my mum is still on :-) For a brief time I offered a lifetime subscription for $100 and sold a few but felt uncomfortable offering to keep the service running forever with no further payments so killed it.

I ended up at $5 / month or $48 / year. It's possible I might increase it a bit more in future but for now I think it's better to keep these prices and try to grow user numbers instead.

One thing I noticed with some companies was that the yearly plan was very heavily discounted compared to the monthly one. To me that gave the impression of lacking confidence that monthly users will stick around long term, so I didn't want to do that.

Yeah I think it's true that popular mainstream consumer products tend to be cheap and highly polished, and that more niche products tend to be higher priced and less polished.

The yearly plan is discounted because it allows self-funding and retains control, versus VC dilution. Explained here: https://youtube.com/watch?v=otbnC2zE2rw (I am not working at present, but I rewatch this every once-in-a-while because it is an amazing talk!)

> One thing I noticed with some companies was that the yearly plan was very heavily discounted compared to the monthly one. To me that gave the impression of lacking confidence that monthly users will stick around long term, so I didn't want to do that.

It probably reflects reality, but I can see the point of showing lack of confidence in the product.

Not a language product, but Bumble (the dating app) at one point -- no idea if they still do it -- offered a lifetime membership priced at around 6 months of monthly subscriptions. I find looking for that kind of pricing can be highly instructive about a business and their competitive landscape, because presumably they'll have run their numbers and are expecting to earn more that way.

So it's not just signalling lack of confidence, but also giving away a very strong signal about the lifetime value of their customers.

At one point my gym offered (out of the blue and only by text) lifetime membership for something like a year or two of the regular monthly price.

They folded nationwide less than ten months later.

That's a good exception - if someone put of the blue appears to push a big discount, it can reflect a high cost of raising capital.

I think it slightly works different with a dating app. In theory, if the app works you don’t need it anymore. So in their case, it is confident to offer a big discount.

In terms of the specific, very low, multiple you're of course right - a dating app is an extreme.

But the overall principle that you should expect the discount to reflect their estimated lifetime value of a customer unless the amount is high enough that the investment return can exceed the regular income. With the exception, as noted by conradfr, that if someone suddenly offers a steep discount it could reflect an attempt to raise expensive capital fast

Assuming you're aware of those factors, it's a useful way of gaining insight into customer retention (assuming they're competent enough to run the numbers...)

Highly polished it depends, but mainstream products certainly won't cater to the long tail of niches. Some niches are still large enough to be profitable for a smaller product, others aren't (and will hardly be supported by any product at all)

Does your ChatGPT thing just asks for the definition of the word or does it look at the context of a few paragraphs around i?

For now it provides the whole sentence that the word was in for context. It works pretty well but I'm sure it would be better if I gave it a little more. 2 or 3 sentences would probably be great, but obviously more expensive to run.

> they asked if I wanted Readlang back > and how much I’d pay. > Of course I said yes!

So how much did you pay?:-)

So was acquired with about 5K MRR?

Getting close to 5k yeah. The average over the preceeding year was around 4k.